pathml

Load whole-slide images across over 160 formats including SVS and DICOM, apply stain normalization, and train deep learning models like HoVer-Net for nucleus segmentation. The framework uses HDF5 to manage large-scale data extraction and spatial graph construction efficiently.

Spatial analysis on CODEX slides is handled by loading raw multiplex immunofluorescence images, performing cell-type annotation, and executing spatial neighborhood enrichment testing. This toolkit provides a unified workflow from image loading to spatial proteomics analysis.

Yes, you can use PyTorch and HDF5 for scalable computational pathology workflows. The framework integrates PyTorch for deep learning-based feature extraction and HDF5 for managing large whole-slide image data, ensuring efficient spatial analysis and model training.

Yes, this computational pathology framework supports proprietary slide formats like NDPI. It seamlessly loads and processes over 160 proprietary whole-slide image formats, including SVS, NDPI, and DICOM, for diverse research workflows.

Build spatial cell and tissue graphs from histology images by loading whole-slide data and using the toolkit's spatial graph construction features. This enables complex spatial neighborhood enrichment testing and downstream spatial proteomics analysis.

You should not use this toolkit if your whole-slide image processing workflow lacks the necessary dependencies or advanced computational resources. It requires PyTorch, HDF5, and Dask to handle large-scale spatial analysis and deep learning model training.